I studied APL before applying for a job where it was used. I was kind of relieved not to get the job. I read somewhere that APL was difficult to write and impossible to maintain. It is so cryptic that it does a while to unravel what the original programmer was thinking. And then those weird keyboards....
The main benefit of APL is programmer job security until a company wisely chucks it!
yeah I wouldn’t want to maintain any of the sample code as I worked through some of the j examples. However after getting back into Linear Algebra and Applied Maths I know what he was trying to do and why. As I try to build more complicated distributed real time systems,(the robot army) the direct connection between the applied math and modern control modeling techniques becomes more attractive. The learning curve is incredibly intense to make things so basic. Haskell has made a nice attempt of bridging the pattern (math) / software impedance mismatch but who the heck is going to use that (I find lots of financial predictive analytics libs/apps with it). Maybe this Rust will pull it all together but in the mean time I’m finding C++ way nicer than I remember it and oh so much better than any of that php,perl,ruby,python, crap that we’ve had to sling during the obama years. Its the cloud everythings or embedded (who doesn’t use os neutral libraries if they can) the same architecture everywhere - compiled compiled compiled. Type safety to important to my sleep to give up. - What version of python does this script run on - 2.7 - of course
A buddy of mine in college was a physics grad student, and programmed in APL. He told me that if an APL program didn't work, and it was 3 lines or longer, it would be less trouble to start over than to try debugging it.
And yes, there was a dedicated keyboard he used!
Mark